1. Field of the Invention
The present invention relates to a method and nodes for performing aggregation of data traffic by relying on Unicast messages in an environment using service bindings over an access domain.
2. Description of the Related Art
Recent years have seen the explosion of Internet Protocol (IP) networks. Initially developed to allow universities and researchers to communicate and cooperate in research projects, it has grown into networks offered at a mass-market level. Nowadays, it is normal for households to have a connection to an IP network to surf the world-wide-web, play interactive games, carry Voice over IP, download documents and softwares, make electronic business transactions, etc.
Reference is now made to FIG. 1, which represents a prior art example of an IP network 100. Typically, an IP network is composed of an access domain 115, network service provider domains 140 and application service provider domains 150. The access domain 115 includes Access Nodes (AN) 120 and an access network 130, such as an IP network. The ANs 120 are network providers, which can offer access to the IP network 130 to user domains 110. The user domains 110 include for example User Devices (UDs) (such as computers, mobile phones, personal digital assistants, etc.), Local Area Networks (LANs) and Wireless-LANs (W-LANs). The user domains communicate with the ANs over various possible technologies. Amongst those technologies can be found dial-up connections and Asymmetric Distribution Subscriber Line connections over telephone lines, cable modems connecting over television cable networks, or wireless communications. The access network 130 is composed of a group of independent routers, which task is to route incoming data traffic based on a destination address embedded therein. As for the network service provider domains 140, they may correspond for example to Voice over IP services, while the application service provider domains 150 may correspond to electronic banking and electronic business transactions.
Though FIG. 1 depicts three user domains, two Access Nodes, two service provider domains and two application service domains, IP networks 100 typically include several thousands of user domains, tenths of Access Nodes, hundreds of service provider domains and application service provider domains. As to the access network 130, it is common to encounter networks including hundreds of routers. It is thus understood that FIG. 1 depicts a highly simplified IP network 100 for clarity purposes.
To ensure a coordinated exchange of data traffic and messages over such IP networks, the IP protocol was developed in the early 1970's. The IP version 4 (IPv4) is used by a majority of currently deployed IP networks. IPv4 provisions for an addressing scheme using 32 bits, which results in a 4,294,967, 296 possible addresses, where each address is unique, and directly identifies one device. In the case of IP networks 100 such as the one shown on FIG. 1, it is commonly known that such network rely on Ethernet-based data link to provide fast and simple transfer of data traffic and messages throughout the IP network 100.
But with the increasing number of devices communicating over the IP networks, and some inherent limitations of IPv4, the IP community has seen the need for a new revision of IP: IP version 6 (IPv6). That new version relies on an addressing scheme using 128 bits, which provides for a much wider number of possible addresses.
Though IPv6 allows for a much greater number of IP addresses, and also addresses some deficiencies found in IPv4, both IPv4 and IPv6 are “best-effort” protocols. “Best-effort” means that a network delivers data traffic without making particular effort to meet higher or particular demands on a quality of service required for those types of data traffic. This might be sufficient for some network service providers 140 and application service providers 150, but unfortunately it proves to be insufficient for others. Thus, some network service providers 140 and application service providers 150 cannot easily and fluidly offer their services over IP networks 100.
To overcome this problem, the MultiProtocol Label Switching (MPLS) is being used over IP networks. MPLS relies on protocols such as ReSerVation Protocol (RSVP) for reserving a path, with a specific quality of service, over the IP network 100. RSVP initially creates a path through a series of routers. To create the path, each router adds an entry to its MPLS table. That entry indicates for data traffic arriving at a specific entry port and having a predetermined label, a corresponding output port and label to be used. By creating such reserved paths in the IP network 100, it makes it possible to carry data traffic for a larger spectrum of network service providers 140 and application service providers 150.
However, with the increasing number of network service providers 140 and application service providers 150 requiring higher quality of service than “best effort”, along with an expansion of the number User Domains 110 and Access Nodes 120 required to allow these User Domains 110 the possibility to use the access network 130, MPLS does not prove to be a good option.
The initial principle at the basis of IP networks is to rely on routers, which perform as few and as little operations as possible before routing incoming data traffic towards their final destination. Also, it is a widely recognized concept that “best effort” networks are a trade-off between quality of service and quantity of data traffic. An increased quality of service, for the same number of routers results in a lower quantity of data traffic being transported on those routers. IP networks have not been designed bearing in mind higher level of quality of service. Thus, by creating reserved paths for higher quality of service data traffic over IP networks, a direct consequence is a reduced quantity of data traffic over those IP networks. In addition, such reserved paths needed for MPLS result in consuming more routing effort in each of the router on the reserved paths. Such routing effort is not significant when only a few reserved paths are open simultaneously, but with the current development of services applications requiring more than “best effort” quality of service, it is possible to envision that thousands of reserved paths will be required simultaneously over the IP networks. Maintaining and routing data traffic with so many reserved paths will become more cumbersome for routers, thus also resulting in slowing routing capabilities of the affected routers. Therefore, the current use of MPLS over IP networks for improving quality of service is resulting in less data traffic being exchanged, and in slower data traffic. Such impacts are not acceptable, as they directly affect all data traffic that is not part of the reserved paths.
There is currently no known solution to the problems associated with the explosion of the number of user devices and of service providers offering services on IP networks. Furthermore, no long-term solution has been identified to allow a tangible and non-destructive solution to the need of increased QoS for certain services and applications.
Accordingly, it should be readily appreciated that in order to overcome the deficiencies and shortcomings of the existing solutions, it would be advantageous to have a method and nodes for efficiently aggregating data traffic using service bindings over access domains, and more particularly a method and nodes for aggregating data traffic using Unicast messages is needed. The present invention provides such a method and nodes.